« Markets in everything | Main | A Lot to Lose »
The AD-AS model with a vertical AD curve
Paul Krugman creates a simple model of the 1930s, based on a vertical AD curve; Greg Mankiw comments and Krugman adds some explanation.
I have a few comments (under the fold)...
1. In the 1940s, Franco Modigliani showed that this kind of argument required infinite liquidity preference, not just very strong liquidity preference. This kind of discontinuity matters in most models with liquidity preference.
2. The model suggests that negative supply shocks don't hurt the economy yet the Dust Bowl was bad for aggregate output. Arguably the AD curve should be kinked and vertical only after some inflection point. But then we are back to negative supply shocks mattering.
3. The model implies that labor unions won't hurt output but the model does not show that labor unions won't hurt employment and indeed the latter question was the original point of contention. A sufficiently high legally binding minimum wage will cause employers to lay off workers, vertical AD curve or not. Maybe capital substitutes for labor or Y stays put for some other reason but employment will go down.
4. I no longer understand Krugman's overall story. He notes that the New Deal years saw a good bit of recovery (I agree), he notes that fiscal policy was not very expansionary (I agree), he believes there was a liquidity trap (I don't agree), and he believes that positive supply shocks run up against a more or less vertical AD curve and thus don't help output (I don't agree). Given all of Krugman's views, AS doesn't much matter and AD didn't much expand. So what exactly drove the (partial) recovery of the 1930s?
Maybe I am taking the model too literally but without the vertical AD curve there is not much else in the model to interpret.
Posted by Tyler Cowen on December 4, 2008 at 07:13 AM in Economics | Permalink
Comments
Krugman also assumes away any real balance effect and then treats it like a fact. Where is the evidence? Surely there was some real balance effect at work and if so, the there would be some downward slope to the AD as noted here.
Posted by: David Beckworth at Dec 4, 2008 7:33:47 AM
I do not feel qualified to have an opinion on a model that a guy like Krugman endorses or on criticisms of that model made by someone like Tyler; however, wouldn't it great if an economist just once responded to a criticism like this with "Yeah, I guess that was kind of dumb." At least one of you has to be wrong. Seriously, fess up.
Posted by: josh at Dec 4, 2008 7:34:48 AM
I read the krugman post and too was confused. I think he was saying something more along the lines of "proposition X has an embedded assumption of a downward sloping AD curve" therefore “if there is no downward sloping AD curve then proposition X may not hold.” He can’t really say things like “I have good evidence that the AD curve was not downward sloping” – instead he says “what if it was vertical? Then what? Hmm? Then…” (He stops short of saying that the AD curve is vertical.)
I’d hazard a guess – based on dynamic system theory, I’m not an economist – that when the slope of a quantity that you generally study becomes zero, the model you want to be looking at is not the model with the zero slope – some other factor becomes more important. You can have discontinuities, etc. Your second derivative becomes important, as does knowing whether you are at an inflection or a min/maxima. So it is problematic to look at this kind of static model at all.
Posted by: babar at Dec 4, 2008 8:36:44 AM
A physicist, chemist and economist are stranded on an island with a can of food. The physicist says if we drop the can from a certain height the impact will open the can. The chemist says we can use the reflection of the sun to heat the lid of the can and it will open. The economist says ASSUME can opener.
Posted by: Jay at Dec 4, 2008 9:12:15 AM
Great link, Tyler! This is a terrific exchange between two great minds both making cogent points quite clearly.
Posted by: steve at Dec 4, 2008 9:51:43 AM
"he believes there was a liquidity trap (I don't agree)"
Tyler,
I would really appreciate an expansion of your disagreement on that point. I think Krugman's argument that Japan suffered a liquidity trap is compelling, I'm less confident of his current claim that the U.S. is in a liquidity trap, and I'd like to know where you think he's wrong on the claim of a trap during the Depression. Is he wrong about Japan, too?
Posted by: James Hanley at Dec 4, 2008 11:10:11 AM
The important thing -- important! -- is to hold on to this "model" and this picture of how to do macroeconomic "science".
Nothing else really matters, does it?
(I mean, besides the excuse this brilliant "rocket science" gives for Krugman's politics).
Posted by: Greg Ransom at Dec 4, 2008 12:37:59 PM
Krugman's story is that cost-push inflation, by which unions push up wages and firms pass on the costs by raising prices doesn't reduce real expenditure and output in a liquidty trap world.
Normally, the higher price level raises the demand for nominal money balances. One effect is for people too sell off bonds to come up with the needed money. This raises interest rates and the higher nominal and real interest rates result in less real expenditure (less investment and consumption.) Firms sell less, produce less, and there is less employment. Another effect is that the higher price level reduces real balances, makes people poorer, and reduces consumption.
In a liquidity trap, those needing more money because prices are higher may sell bonds, but those are bought by the bear speculators who are keeping interest rates from falling. Basically, bond specultors reduce their money hoards, freeing up money to pay the higher prices of goods and services at unchanged levels of output.
Krugman, so far anyway, hasn't said much about bear speculators but is rather talking about the "zero" bound on nominal interest rates. As far as I can tell, he is pretty vague about what happens exactly in that scenario. Since his liquidity trap involves the central bank standing ready to create extra money at the zero interest rate, then I suppose that what happens is that the Fed creates enough new money to finance the current level of real expenditure at the higher price level. The Fed cannot get interest rates lower by expanding the money supply (they are at zero,) but if there is an increase in the demand for money, because of higher prices, then the Fed can accomdodate that, and keep interest rates zero.
Tyler's response is that this model seems inconsistent with some real drop in productive capacity causing output to fall. A drought.
In a depression, the capacity constraint, which would be represented by a vertical aggregate supply curve would be to the right of the supposedly vertical aggregate demand curve. In a one good economy, anyway, as long as the aggregate supply curve shifts no more left than the aggregate demand curve, the drought has no impact on output.
If there really are farm surpluses.. then the drought would just reduce them. The farmers in the drought striken midwest would inspire novels. The farmers in the east whose can now sell all of their crops and are doing OK in the depression.. I guess no one pays them much mind.
If we really had a situation where people couldn't get food because of the drought, and they didn't want to buy anything else with the money they cannot spend on the nonexistent food, then aggregate demand
would sift to the left. If on the other hand, they buy something else with the money, then the composition of output shifts from food to whatever it is they want to buy, but aggregate output remains the same.
Maybe one of those AS curves that look a bit like a reverse L would be appropirate. Krugman is talking about the bottom of the L shifting up. Cowen is talking about the top of the L shifting left.
Cowan, of course, is claiming that as an a matter of fact, the "Dust Bowl" did have an adverse impact on the economy. How about the farmers are spending less, so aggregate demand falls? (The farmers who sell more just save?)
My opinion is that the Fed failed to buy up the entire national debt and didn't even start on created ways to expand the money supply. Just saying, we won't bother because it won't do any good because nominal interest rates are low by historical standards--well, that doesn't cut it. If they try, and the money supply rises enough to offset the drop in income velocity, and income velocity just drops even more. And their is nothing else for the central bank to buy... well, then I would start to believe.
Still, I think a currency suspension and negative nominal interest rates might be a good approach sometimes.
Posted by: Bill Woolsey at Dec 4, 2008 1:10:28 PM
I've also found Krugman to be a bit incoherent, arguing positions counter to his own positions in a misdirected effort to debunk other positions that would be better debunked using other arguments.
But condensed, I think (?) he'd say:
The *net* effect of policies '32-'36 was mildly expansionary but overall rather tepid, insufficient to break the negative cycle and re-establish a virtuous equilibrium.
Net policy effects '36/'37 were contractionary, causing a resurgence of the still-lurking depression.
European orders starting in '39 (yes, Tyler, I think he'd give you that one) followed by massive federal deficit spending beginning '41 finally broke the depression's back, creating the necessary demand to bring production back up to capacity.
And I'd say: True, consumer demand didn't really increase during the war years, for reasons unrelated to fiscal or monetary policy (i.e., the war). But savings (E bonds especially) did increase. Once the war's distortionary effects were removed, with the wheels of the economy spinning, the great prosperity arrived.
Posted by: Steve Roth at Dec 4, 2008 1:56:07 PM
I've never been able to understand what a liquidity trap is. And therefore Krugman's vertical AD makes little sense to me.
I propose the following: A liquidity trap is an artifact of thinking about the world as having a single interest rate. Once you have "liquidity trap" conditions (i.e. the economy is at risk of Depression) interest rate correlations go beserk -- and we are very far from a single interest rate representing the real world.
Maybe what is called a liquidity trap is really just a bank funding crisis. Banks stop lending because they face a crisis in raising funds to lend. It's possible that this problem can be solved by a combination of well-targeted government support and the exploitation of interest rate differentials. (This is precisely what the Fed is doing. I tend to think that Japanese banks should have offered higher interest on deposits -- but that also implies that interest rates paid by firms would have to be higher -- did commercial paper markets or their equivalent undermine the Japanese banking system?)
Because of the delay in adjusting to a new bank funding equilibrium, many credit dependent aspects of the economy slow or stop. And there is a definite need for government help to bridge these difficulties -- money transfers to the poor and unemployed strike me as the most sensible way to address this, because the money will be spent. I don't have a problem with financing long-overdue infrastructure projects.
But beyond the fact that government support is best spent where it will be recycled back to the economy (i.e. not saved), I don’t get the value of the aggregate demand framework. What we have is a banking crisis that is affecting the real economy. Any solution will require dealing with the fact that long-standing interest rate relationships have gone haywire (classic evidence of financial disarray). We can be pretty sure of one thing: the crisis will only be over when interest rates start moving together again.
Posted by: ccm at Dec 4, 2008 3:05:32 PM
Here's what confuses me. Assuming the existence of a liquidity trap, I thought that means that there's a floor to the interest rate, not a ceiling. So the aggregate demand curve would only become vertical at some point to the right, not for its entire length. If Krugman's argument were right, there would be no limit to how high you could push the price level relative to the money supply without affecting output. Or to put it more simply, you could legislate a minimum wage as high as you please without affecting employment. Which doesn't make sense. Or am I missing something?
Posted by: Phil P at Dec 4, 2008 5:25:16 PM
He assumes competitive labor markets, as if labor unions didn't exist, causing unemployment to be higher than it would have been with competitive labor markets. Amity Shlaes doesn't assume this.
And the concepts of aggregate demand and aggregate supply are bogus. How do you "aggregate" the demand/supply for wheat, with the demand/supply for medical services, with the demand/supply for cars, with the demand/supply for teachers services, etc.? You can't.
Posted by: Bill Stepp at Dec 4, 2008 6:02:24 PM
Btw, as Alan Meltzer has pointed out (see his homepage for instance), no one has ever seen a liquidity trap.
A certain helicopter pilot has pointed out that there is a technology called a printing press, as well as stuff called paper and ink that the Fed could use to cause inflation. No open market ops required. Back to Meltzer: at least some of the new money would be spent. So much for his post.
Posted by: Bill Stepp at Dec 4, 2008 6:07:01 PM
Money is the primary determinant of aggregate spending.
The demand for money is sensitive to interest rates, but not enough to cause a liquidity trap. In a period of low interest rates banks make lower interest loans but also pay lower interest rates to depositors. The spread on interest rates stays about the same over time.
The less interest rate sensitive money demand is, the more effective monetary policy is Vs fiscal policy.
Even if interest rates are near zero the Fed can use other instruments then short term government securities. The Fed can buy foreign currencies or example.
I think Phil P is correct. If you take Krugman's argument to the extreme, we should just triple the minimum wage and the economy will improve.
However I think Krugman may mean that the AD curve is vertical for a period of time over some relevant range of prices and output. How you determine that in such a way that you can implement government programs to correct for it in a meaningful way, given political constraints, time constraints etc. is unclear. I would go back to the Milton Friedman argument that by the time politicians identfy the problem free markets are solving it.
Plus if you want real long term growth you want to shift the AS curve to the right, which can be down with saner tax and regulatory policies.
Posted by: DanC at Dec 4, 2008 8:05:38 PM
BTW
Did Milton say something nasty to Krugman that Krugman has such an ax to grind?
Posted by: DanC at Dec 4, 2008 8:06:54 PM
Even if the great depression did have a period when the creation of private cartels helped to assure profits and some expansion of output (with Congress demanding increased wages for workers in return for allowing the cartels) it is far from clear that we now have the idle manufacturing capacity that was present then. So even if I agreed with Krugman about the vertical AD, how does Krugman separate the effects of the cartel from the increased wages of the new deal?
Our current crisis had two steps. An oil price shock started a mild recession. The Fed seemed worried about inflation. Then a bubble of creatively created mortgages started to reset. Suddenly we saw that financial firms were taking on a great deal more risk then we realized. The core crisis is the huge drop in home values and the impact on the financial sector.
Where is the excess capacity in the core housing issue? Or do we just raise wages until they are inline with inflated housing prices?
How the fiscal stimulus that Krugman proposes deals with the core housing issue is unclear. And why he thinks that the government is better at finding investment opportunities then the private sector is also unclear. It would seem that Krugman wants BIG government and is using the current crisis to push an agenda that he has regardless of the economy.
Posted by: DanC at Dec 5, 2008 2:09:25 AM
The time frame of the model is important. My understanding is that the model is for the next 6 months or so, perhaps one year. In this short time you would not expect capacity to be affected much on the capital side, and a little on the labor side (through layoffs, though existing labor presumably will be working that much harder to offset the loss of labor force in production units that don't actually close down).
I too would like to know: when do we know for sure there is or isn't a liquidity trap?
Oh, and one last comment particularly addressed to non-economists: these curves are "expected" curves in the sense that they relate a real quantity (real output) to a price, and that makes a not-too-straightforward interpretation: my understanding (correct me if I'm wrong) is that for every jerk in expectations you'd be jumping about the plane, presumably not necessarily locally but potentially all over the place.
Posted by: pat toche at Dec 5, 2008 12:39:01 PM
Our current crisis had two steps. An oil price shock started a mild recession. The Fed seemed worried about inflation. Then a bubble of creatively created mortgages started to reset. Suddenly we saw that financial firms were taking on a great deal more risk then we realized. The core crisis is the huge drop in home values and the impact on the financial sector.
DanC, I believe you're looking at the wrong questions here.
Sure, the financial sector was basicly running a ponzi scheme on housing, and they got caught with a lot of inventory themselves that they hadn't passed on to customers. (Or maybe the real ponzi operators got out fine and it's bankers who got sucked in who were caught.) And sure, the little oil increase made some difference. But there were fundamental problems which were far more central.
The US dollar was not sustainable as a reserve currency, but no one had a good alternative so they limped along with it.
The US/europe balance of trade against east asia needed some sort of resolution, but there was no sign of one.
USA used 1/4 of the diminishing oil, did not produce exports to pay for it. Who should get the oil, the USA or china which produces stuff the USA wants to buy?
Every nation has a comparative advantage. The USA's comparative advantage was for complex financial derivatives. Demand for these products is low just now; how will we find other comparative advantages so we can produce value to the global economy to match what we consume? Or is that the wrong question?
It's like, when month after month you spend more money from your checking account than you put in, sooner or later you're going to bounce some checks. When you do, it isn't so interesting precisely which checks you bounced.
Posted by: J Thomas at Dec 5, 2008 10:30:14 PM
Our current crisis had two steps. An oil price shock started a mild recession. The Fed seemed worried about inflation. Then a bubble of creatively created mortgages started to reset. Suddenly we saw that financial firms were taking on a great deal more risk then we realized. The core crisis is the huge drop in home values and the impact on the financial sector.
DanC, I believe you're looking at the wrong questions here.
Sure, the financial sector was basicly running a ponzi scheme on housing, and they got caught with a lot of inventory themselves that they hadn't passed on to customers. (Or maybe the real ponzi operators got out fine and it's bankers who got sucked in who were caught.) And sure, the little oil increase made some difference. But there were fundamental problems which were far more central.
The US dollar was not sustainable as a reserve currency, but no one had a good alternative so they limped along with it.
The US/europe balance of trade against east asia needed some sort of resolution, but there was no sign of one.
USA used 1/4 of the diminishing oil, did not produce exports to pay for it. Who should get the oil, the USA or china which produces stuff the USA wants to buy?
Every nation has a comparative advantage. The USA's comparative advantage was for complex financial derivatives. Demand for these products is low just now; how will we find other comparative advantages so we can produce value to the global economy to match what we consume? Or is that the wrong question?
It's like, when month after month you spend more money from your checking account than you put in, sooner or later you're going to bounce some checks. When you do, it isn't so interesting precisely which checks you bounced.
Posted by: J Thomas at Dec 5, 2008 10:30:44 PM
"The Fed can print money and cause inflation, dropping it from the air," is poorly thought out. It is true.. but poorly thought out.
First of all, printing up money and giving it away combines fiscal and monetary policy. It is like a tax rebate and an increase in the quantity of money.
In basic Keynesian economics, the IS and LM curve shift to the right. The aggregate demand curve shifts to the right. This applies in both the the liquidty trap and the more usual situation. The LM curve is horizontal in a liquidty trap (which is usually drawn as a range,) and positively sloped "normally." IS is negatively sloped, showing the negative relationship between real interest rates and real expenditures.
The way this would really be done is to have a tax rebate, financed by Treasury borrowing--bond sales. The central bank purchases the bonds. New money created out of thin air is given to people to spend.
I suppose distributing tax rebates in some kind of random way similar to picking up money dropped from the sky might make a difference, but not much regarding aggregate demand.
Another issue, covered by Krugman, has to do with "responsiblity." If the goal of the central bank is to temporarly expand the money supply in order to increase aggregate demand and maintain some kind of target for the growth path of nominal expenditures or the prices level, then, they have to stand ready to wihdraw the money back out of circulation.
This is different from a situation in which newly created money permanently increases the equilibrium price level and the central bank just doesn't care. Then there is no implicit commitment to withdraw the money from ciculation when it is no longer "needed."
In the realistic scenario, the central bank needs to stand ready to sell off the bonds that it purchased from the central bank.
And so, the tax rebates do increase the national debt and create future tax liabilities, even if they are combined with an expansionary monetary policy.
And the same is true of government spending projects, even if the fiscal policy is supported by an expansionary monetary policy.
If dropping newly printed money from the sky increases spending, if it not to cause inflation, there has to be an implicit committment to withdraw the money from circulations--mostly simply by selling government bonds.
The effect of the fiscal policy element of printing currency and dropping it from the air is the standard question regarding whether people will spend or save their tax rebates. The usual assumption is that they will spend some of it, but not all.
And so, even in a liquidity trap, printing money and giving it away will expand spending in the economy.
To the degree that people save and lend the money they pick up, in a liquidity trap, this will not lower interest rates, because bear speculators will sell off bonds and accumulate all the addional money. (Patinkin once pointed out that the bear speculators don't really have infinite amounts of bonds to sell, and so, they can be "defeated" by the central bank if it is willing to print up money and give it away.)
The liquidity trap does suggest that fiscal policy alone can expand aggregate demand just as effectively as the combined fiscal policy/monetary policy above. There is no need for the central bank to increase the money supply because bear speculators will provide all the money needed at constant interest rates. There is no nominal crowding out because of fiscal policy.
So, the tax rebates would be equally effective whether they were combined with newly created money or if the central bank did nothing and so, they are directly financed by bonds.
The liquidity trap describes a situation in which monetary policy cannot lower interest rates. If lower interest rates are the only way that expanding the money supply increases spending, then, in a liquidity trap, monetary policy alone cannot stimulate aggregate demand.
The liquity trap doesn't say that a combined monetary and fiscal policy (like printing up money and giving it away) cannot impact aggregate demand. It does say, remarkably, that having the govennment borrow money and give it away has an equal impact on aggregate demand as printing it.
The vertical aggregate demand would generally be a "range." At least in the traditional, bear speculator story. If the price level is high enough, the aggregate demand curve would have its ordinary negative slope. Again, this argument ignores he "pigou effect," of changes in the real balances on wealth and consumption. Krugman mentions this and dismisses it as being too small. (a 10% increase in the price level would should increase real wealth by an equivalent of 140b, if base money is fully real wealth in our current monetary regime. With a gold standard, a change in the price level creates real wealth in proportion to the size of the gold stock--a true outside money.)
I don't understand why Krugman says we are in a liquidity trap now. As I have mentioned before, it would seem that he really believes that monetary policy alone cannot expand aggregate demand enough to return to full employment any time soon. And that if the Fed tries, then it will soon be in a liquitidy trap. YOu know, interest rates aren't at zero yet. He seems to be saying that once they get to zero, that will be all the expansionary monetary policy we can have, and so, we will be in a liquidty trap. And, further, that aggregate demand will still be too low.
There is some (well, lots) of evidence that the Fed is not going to follow a policy of saying that-- well, the federal fund rate and 3 month T-bills are near zero, so, we have done all we can do. That they will seek to pull down other interest rates (by purchasing other kinds of securities) and so, the traditional differene between, the t-bill rate and other interest rates will be bround down.
So, I suppose, in the extreme BBB corporate bonds will have to be zero along with 3 month T-bills.
And, of course, they appear to realize that interest rate targetting can create a disaster, especially when there are deflationary expectations. They will have to look at the money mutipliers and measures of the money supply, and not gauge monetary policy by interest rates.
Posted by: Bill Woolsey at Dec 6, 2008 1:15:33 PM
